Host polar-angle adapter for executable binary32

Complex logarithms use the host binary64 atan2 and round its result to binary32. Finite inputs embed exactly in binary64. This adapter is a host transcendental boundary; it is not a certified or platform-independent binary32 arctangent implementation.

@[instance_reducible]

instance TorchLean.instAtan2BinaryOfNatNatIeeeDefaultBias : Atan2 (FloatLib.Floats.ExecFloat.Binary 8 23 FloatLib.Floats.Formats.BinaryInterchange.FloatFormat.Encoding.ieee (FloatLib.Floats.Formats.BinaryInterchange.FloatFormat.Encoding.ieee.defaultBias 8) instAtan2BinaryOfNatNatIeeeDefaultBias._proof_1 instAtan2BinaryOfNatNatIeeeDefaultBias._proof_2 instAtan2BinaryOfNatNatIeeeDefaultBias._proof_3 instAtan2BinaryOfNatNatIeeeDefaultBias._proof_4)

Host polar angle rounded from binary64 to executable binary32.